Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts
Abstract
Fast prediction of the response of adhesive soft viscoelastic contacts represents a current challenge in soft robotics and for gripping and manipulation tasks. Determining the complete time-resolved force trajectory requires full numerical simulations, whose computational cost is strongly parameter-dependent, making them impractical for real-time application or design-optimization loops. In this work, we overcome this limitation by training a scalar-conditioned, stateful, sequence-to-sequence deep learning model to predict the full force evolution from a prescribed displacement history for both short- and long-range adhesion regimes. The data set spans four orders of magnitude in loading and unloading rates and includes varied dwell times, with the Tabor parameter ranging from $0.2$ to $3.2$. To enable learning across these heterogeneous time scales, we introduce a fixed-measurement-step (FMS) representation that converts variable-length trajectories into fixed-length sequences while preserving their physical-time information. Different architectures were trained, including long short-term memory (LSTM) networks, temporal convolutional neural (TCN) networks, and time-distributed dense layers with three different Tabor-conditioning mechanisms. The models were compared using global waveform and error metrics. We found that the best-performing model has an LSTM architecture with concatenated conditioning, which achieves a held-out mean-squared error of 5.0×10−4, a median pull-off-force error of ≈2.2%, and a median hysteresis error of ≈1.1%. For the held-out protocols, the model predicts a complete force trajectory with a median inference time of $0.16$ s. The model is tested across unseen parameter combinations and against analytical limiting cases, providing a rapid surrogate for repeated numerical evaluations with potential use in control-oriented applications.
Related
- Mutation-resolved ABL1 inhibitor resistance prediction under mutation-disjoint and temporal cross-study validation
- An automated ranking platform for machine learning regression models for meat spoilage prediction using multi-spectral imaging and metabolic profiling
- Adhesive riffle shuffles: one-parameter models and mixing bounds
- A Prediction--Correction Analysis of Two-Way Block Splitting in Distributed Learning
- Note on the Exponential Stability of the Viscoelastic Wave Equation
BibTeX
@misc{maghami2026deep,
title = {{Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts}},
author = {Ali Maghami and Merten Stender and Michele Ciavarella and Antonio Papangelo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2607.19060},
eprint = {2607.19060},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.19060},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts
AU - Ali Maghami
AU - Merten Stender
AU - Michele Ciavarella
AU - Antonio Papangelo
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2607.19060
UR - https://arxiv.org/abs/2607.19060
AB - Fast prediction of the response of adhesive soft viscoelastic contacts represents a current challenge in soft robotics and for gripping and manipulation tasks. Determining the complete time-resolved force trajectory requires full numerical simulations, whose computational cost is strongly parameter-dependent, making them impractical for real-time application or design-optimization loops. In this work, we overcome this limitation by training a scalar-conditioned, stateful, sequence-to-sequence deep learning model to predict the full force evolution from a prescribed displacement history for both short- and long-range adhesion regimes. The data set spans four orders of magnitude in loading and unloading rates and includes varied dwell times, with the Tabor parameter ranging from $0.2$ to $3.2$. To enable learning across these heterogeneous time scales, we introduce a fixed-measurement-step (FMS) representation that converts variable-length trajectories into fixed-length sequences while preserving their physical-time information. Different architectures were trained, including long short-term memory (LSTM) networks, temporal convolutional neural (TCN) networks, and time-distributed dense layers with three different Tabor-conditioning mechanisms. The models were compared using global waveform and error metrics. We found that the best-performing model has an LSTM architecture with concatenated conditioning, which achieves a held-out mean-squared error of 5.0×10^−4, a median pull-off-force error of ≈2.2%, and a median hysteresis error of ≈1.1%. For the held-out protocols, the model predicts a complete force trajectory with a median inference time of $0.16$ s. The model is tested across unseen parameter combinations and against analytical limiting cases, providing a rapid surrogate for repeated numerical evaluations with potential use in control-oriented applications.
ER -